The Nomad Rover Was Deployed for 45 Days in the Atacama Desert, Chile, During the Summer of 1997. During This Period, the Rover

Lunar and Planetary Science XXIX 1013.pdf

ATACAMA I: SCIENCE RESULTS OF THE 1997 NOMAD ROVER FIELD TEST IN THE ATACAMA DESERT, CHILE. N. A. Cabrol (Project Lead Science Investigator), G. Chong Diaz2 (Field Science Team coordinator), Science Team: J. Dohm3, M. Pereira Arredondo2, G. Dunfield1, V. Gulick1,4, A. Jensen Iglesia 2, R. Keaten5, C. Herrera Lamelli2, R. Landheim1,4, P. Lee1, L. Pedersen6, T. Roush 1 , K. Schwher7, C. Stoker1, A. Zent1. 1Space Science Division, NASA Ames Research Center, Mof- fett Field CA 94035, USA, 2Universidad Catolica del Norte, Antofagasta, Chile, 3U.S. Geological Survey, Flagstaff AZ 86001, USA, 4SETI Institute, NASA Ames Research Center, Moffett Field CA, USA, 5U.S. Geological Survey, Menlo Park CA, USA, 6FRC, Carnegie Mellon University, Pittsburgh PA, USA, 7IMG, NASA Ames Research Center, Moffett Field CA, USA ([email protected].nasa.gov).

The Nomad rover was deployed for 45 days in the desert experience for remote operators through high- Atacama Desert, Chile, during the summer of 1997. quality imagery, (2) simulate NASA missions (Mars, During this period, the rover set the record of the long- Moon, and Antarctica) by: (a) training scientists, est traverse ever performed by an automated vehicle (b) evaluating control environment appropriateness, (220 km), while controlled by operators either at (c) developing and evaluating exploration strategies, NASA Ames and Carnegie Mellon. During this trav- (d) evaluating best interaction by distributed science erse, between June 20th and 27th, the rover was used teams, (3) evaluate the importance of various image to perform science experiments. Both Science and Op- techniques, i.e., panospheric imaging, pan/tilt camera, eration Team controlling the rover were located at stereo imaging, close-up imagery, (4) understand the NASA Ames, while a Field Science Team was in the reasons of data misinterpretation in previous field ex- Chilean Desert to ground-truth the operations. periments by ground-truthing with feedback to the Sci- ence Team at NASA Ames, and careful evaluation of scientists’ procedures and protocols. Nomad Rover: Capabilities and Science Pack- age: Nomad is a four-wheel drive, four-wheel steer robust rover (wheel diameter: 76.2 cm), of 2.4 x 2.4 x 2.4 m deployed, and 550 kg mass, with a transformable chassis. The rover used an actively pointed antenna to support high-bandwidth telemetry [1], [2]. Nomad’s average speed was 0.3 m/sec, though it was able to sustain 0.5 m/sec during the simulated Mars and Moon operations in open terrain. Navigation was enabled by Inertia Measurement Unit, Gyrocompass, and Global Positioning System (GPS), with a precision of about one meter. The imagery system included a panoramic The Atacama Desert: The Atacama desert in- camera with an ultrawide field of view (360° x 40° cludes a variety of features and characteristics that above horizontal) using a spherical mirror mounted make this site a unique place to perform experiments above the vertically oriented digital camera [3]. The on planetary-analog surfaces (i.e., Mars and the science imaging system (high resolution cameras) was Moon): craters from meteoritic impacts, volcanoes, designed to reach the resolution of the human eye (see rocks and sand fields, total lack of vegetation due to specifications below): the absence of rain (1 cm/year from fog), ancient - and • Color Stereo Cameras: 640x480 pxl; 0.29865 now dry - lake beds and sea floor. The average eleva- mrad/pxl; 8 x 11° FoV; 8 bit per color (24 total); tion of the test area (23°20’N/68°37’W) was 2400 m, left/right stereo separation: 25 cm; and the diurnal temperature variation from 0° to 25°C. • Black and White Stereo Cameras: 640x480 pxl; The terrain offered challenges to test the vehicle mo- 0.895 mrad/pxl; 25 x 33° full FoV; left/right stereo bility and trafficability capacities, with a succession of separation: 12.5 cm; soft surfaces due to the accumulation of smooth mate- • Panospheric Camera: 1024 x 1024 pxl; 360° FoV; rials, and ravines left by ancient channels currently dry 1 Mbyte per image; which eroded the desert in the past. • Science Instrument: Magnetometer: sled dragged Overall Project and Science Objectives: The behind Nomad on rigid towbar; sensor footprint: 8” at field test was designed to demonstrate and validate: (a) 13 cm. robust locomotion, navigation, visualization, and • Aerial Photos: 1 m/pxl resolution communication in a long-distance, long-duration trav- • Weather Sensors: Temperature; humidity, wind erse in Chile’s Atacama desert, (b) perform end-to-end velocity. trial under remote control; teleoperation and autono- Exploration Strategies: The goal was to prepare mous control with simulated time delays. near-term planetary missions by testing different ex- The objectives pertaining to the NASA Ames sci- ploration strategies. The planetary mission scenarios ence field experiment were to: (1) provide realistic included: (1) Mars Caching Samples by a thorough examination of the sites, with and without panospheric Lunar and Planetary Science XXIX 1013.pdf

NOMAD ROVER FIELD TEST: N.A Cabrol and Nomad Science Team

camera, and with and without time delay), (2) Science C. Mina, M. Sims, H. Thomas, D. Tucker, and E. “on the fly”, while keeping the rover in motion 75% of Zbinden, and from the University of Iowa, Lab. for the operation time to determine whether or not suc- development of Telepresence Interface, Dr. G. Tho- cessful interpretation of the environment is related to mas. A special thank to: T. Bunch, R. Mancinelli, J. the time spent on a target, (3) Exploring strategies Farmer (NASA ARC), and D. Lowe (Stanford) for (visual and instrumental) to remotely identify meteor- providing their expertise during the analysis of sample ites in an extreme environment (i.e., in Antarctica). Results: (a) The Science Team at NASA Ames 1-250697. This work was supported by the NASA properly identified the geologic environment, often to Space Telerobotics Program, David Lavery, Program detail, when compared with the ground-truth provided Manager. later in the operation by the notes from the Field Sci- References: [1] Whittaker, R. et al., 1997. ence Team, (b) the science “on the fly” successfully Proceed. Intern. Symp. Artif. Int. Autom. for Space., showed that the selection of appropriate targets might [2] Cabrol, N. A. 1997. Nomad Atacama Desert Trek: be even more critical than the time spent on a study Science Plan (Mars, Moon, and Antartica Simulated area to reconstruct the history of a site. This operation Operations), SST Int. Report (available at also allowed the Nomad Team to set a new record of http://img.arc.nasa.gov), [3] Wettergreen et al., 1997. traverse-length for a one-day operation (1.3 km), and Operating Nomad during the Atacama Desert Trek (in during the same day, a new rock unit was discovered press). by the Science Team using the rover in a local Jurassic outcrop, (c) both visual and instrumental modes dem- onstrated the feasibility, in at least some conditions, of carrying out a field search for meteorites by remote control. Further field tests of a range of techniques, tools and strategies for the identification of meteorites in the field is one of the prime goals of the 1997 No- mad expedition to the Patriot Hills, Antarctica, that is following the Atacama Field experiment. Conclusion: The Atacama Field experiment was able to finally breach the barrier of geologic misinterpretation often encountered in previous field tests. Sedimentary, metamorphic, and igneous rocks were properly identified using only visual instruments, and the main lines of the site’s recent geologic history (~ 250 Myr) were reconstructed. The human-eye resolu- Figure: Nomad on a rocky field heading to the salar tion camera, the panospheric camera providing the (ancient lake now dry). The association of ancient aqueous Science Team the feeling of being in the field by ena- sedimentary deposits, channels, rocky and sandy fields, and bling a 360° observation of the study area (even with volcanoes provided a good analog to planetary missions. The the rover in motion), and terrain models comparable to Andes in the background. those used two weeks later during the Pathfinder mission, were important to the success of the Nomad mission. The individual benefit and value of these instruments and techniques will be assessed in future studies. Acknowledgment: The Nomad project is led by the Carnegie Mellon University (CMU). We thank Dr. R.Whittaker, E. Rollins, M. Maimone, J. Murphy, M. Sibenac, D. Bapna, J. Teza, M. Parris, B. Shamah, and A. Foessel for enabling our participation to the Project. The Atacama field experiment was a joined effort of the CMU Field Robotics Center (FRC), the NASA Ames Research Center Intelligent Mechanism Group (IMG), and the Space Science Division (SST). We are very thankful to our field Science Team in Figure: The succession of ancient dry channels, often as- Chile from the Universidad Catolica del Norte (UCN), sociated with loose material provided challenges for the Antofagasta, for its dedicated collaboration. These navigation team. tests were possible thanks to the tireless collaboration of the IMG Operation Team: D. Wettergreen (IMG Project Operation Manager), M. Bualat, D. Christian,